eval
The segmentation neural network evaluation logic.
kelp.nn.training.eval.EvalConfig
Bases: PredictConfig
Config for running NN model evaluation.
Source code in kelp/nn/training/eval.py
21 22 23 24 25 26 27 28 29 30 31 32 |
|
kelp.nn.training.eval.main
Main entrypoint for model evaluation.
Source code in kelp/nn/training/eval.py
131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 |
|
kelp.nn.training.eval.parse_args
Parse command line arguments.
Returns: An instance of EvalConfig.
Source code in kelp/nn/training/eval.py
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
|
kelp.nn.training.eval.run_eval
Runs model evaluation.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
run_dir |
Path
|
The run directory. |
required |
output_dir |
Path
|
The output directory. |
required |
model_checkpoint |
Path
|
The model checkpoint path. |
required |
use_mlflow |
bool
|
A flag indicating whether to use MLFlow to load the model. |
required |
train_cfg |
TrainConfig
|
The original training config. |
required |
experiment_name |
str
|
The experiment name. |
required |
log_model |
bool
|
A flag indicating whether to log model as an artifact. |
False
|
tta |
bool
|
A flag indicating whether to use TTA. |
False
|
tta_merge_mode |
str
|
TTA merge mode. |
'max'
|
decision_threshold |
Optional[float]
|
An optional decision threshold. Will use :meth: |
None
|
Source code in kelp/nn/training/eval.py
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 |
|